AI Agents in QA
Course Objective
To teach QA professionals how to use AI agents to automate routine testing tasks: from requirements analysis and test case generation to creating and running automated tests, analyzing results, and working with defects.
Participants do not just get acquainted with AI capabilities, but create their own AI-powered QA workflows and QA agents that can perform a sequence of testing tasks with minimal manual intervention.
Who is this course for?
- QA Manual and QA Automation Engineers
- Test Engineers
- QA Leads and Test Leads
- SDET
- Software Engineers involved in testing
- QA professionals who already use ChatGPT, Claude, or Gemini, but want to move from prompting to agentic workflows
- QA Managers who want to understand the capabilities and limitations of AI Agents in QA
Required: experience in testing, writing test cases, bug reports, and a deep understanding of the SDLC. For some automation practices, experience with web testing is required.
Course Outcomes
Upon completion of the course, participants will be able to:
- distinguish between AI Assistant, AI Agent, and Agentic Workflow;
- identify QA tasks that are appropriate to delegate to AI agents;
- use AI for requirements analysis and finding potential defects;
- generate test scenarios, test cases, edge cases, and test data;
- create AI-assisted exploratory testing workflows;
- use AI to generate and maintain automated tests;
- work with browser automation through AI agents;
- use MCP to connect AI agents to QA tools;
- apply Playwright in conjunction with AI agents;
- use AI to analyze failed tests and perform initial root-cause analysis;
- automate the QA workflow from requirement to defect report;
- create their own specialized QA Agent;
- evaluate the quality of AI results and build human-in-the-loop control;
- understand the risks of using AI.
What will you get from the course?
Program
- How the role of QA is changing in the era of Generative AI
- AI-assisted testing vs AI-driven testing
- AI Assistant, AI Agent та Agentic Workflow
- What the agent can do independently: plan → act → observe → analyze → adapt
- Where AI really saves QA time, and where it creates additional risks
- Human-in-the-loop as part of the modern QA process
Practice:
- Analysis of own QA workflow
- Identifying tasks that can be delegated to AI
- Building the first AI-powered QA workflow
- Analysis of requirements, user stories and acceptance criteria
- Detection of ambiguities and missing requirements
- Generation of test scenarios
- Positive / negative / boundary / edge cases
- Risk-based test design using AI
- Traceability між requirements та tests
Practice:
Participants submit an AI user story and receive:
Requirement → Risks → Test Scenarios → Test Cases → Edge Cases
- Synthetic test data
- Limit values
- Invalid data
- Combinations and borderline cases
- Mass variants of test data
- AI as "second tester / second tester"
- Search for unusual scenarios
Practice:
Creating an AI-driven exploratory testing session for a real or demo product.
- Generation of automated tests
- Conversion of manual test cases to automation
- Coding with AI
- Generation of selectors and assertions
- Refactoring of automated tests
- Test service
Practice:
- Creating an automated test using AI
- Generating tests from natural language
- Analysis and improvement of test code
- How the AI agent interacts with the web application
- Browser automation
- Agent perception → planning → action
- Context and tool call
- MCP: what is it and why QA
- MCP servers for testing workflows
- Connecting an AI agent to browser automation
Practice:
Creation of workflow:
AI Agent → Browser → Application → Test Script → Result
Participants give the agent a task in natural language, and the agent independently performs a sequence of actions in the web application.
- Scheduling a test run by an agent
- Choice of tests
- Running tests
- Analysis of results
- Working with failed tests
- Unreliable tests
- Root cause analysis
- Test reports generated by artificial intelligence
Practice:
AI receives test run results and independently performs:
Detection → Analysis → Classification → Explanation → Recommendation
- Agent orchestration
- Single-agent and multi-agent workflows
- Automation of repetitive quality control operations
- What to leave behind a person
- Hallucinations
- False positives / false negatives
- Security and privacy
- Access control
FAQ
The course is intended for QA Automation Engineers, Test Engineers, SDET, QA Leads, Test Leads and other professionals who want to use AI Agents to automate and optimize testing processes. It is important to have basic knowledge of Software Testing, SDLC, test cases and bug reports. For practical work with automation - experience with web testing.
The course goes beyond using ChatGPT or Claude to write test cases. Participants learn to build agentic workflows in which AI can independently perform a sequence of QA tasks: analyze requirements, create test scenarios, work with browser automation, run tests, analyze results and prepare reports.
The program uses modern AI and QA tools, including AI assistants/agents, Playwright, MCP and tools for AI-assisted test automation. A specific set of tools can be adapted to the level of the group and the technology stack of the company.
The participant will be able to use AI Agents to analyze requirements, create test scenarios and test data, exploratory testing, generate and support automated tests, analyze failed tests and prepare bug reports. The main practical result is an own AI-powered QA workflow and QA Agent that can be adapted for use in a real project.
Yes. The program is especially suitable for QA teams that plan to systematically implement AI in testing processes. The corporate version can be adapted to a specific company stack, level of participants and real QA processes: Jira, Confluence, Git, CI/CD, Playwright and other tools.
Training combines short theoretical blocks, demonstrations and practical work. The main emphasis is on participants not only getting acquainted with AI Agents, but also learning how to use them in real QA scenarios.
Yes. For corporate clients, we can adapt practical tasks to real products, QA processes and the technological stack of the team. It is also possible to expand the program to a deeper practical format with the construction of an end-to-end Agentic QA Workflow for a specific company.
Education provides
